Note: Descriptions are shown in the official language in which they were submitted.
CA 02313485 2000-06-09
WO 99/30181 PCT/US98/24788
Method and System for Determininct Position of a Mobile Transmitte_
Background of the Invention
The present invention relates generally to position detection methods and
systems,
and more particularly, to a method and system for detecting a position of a
mobile device
or transmitter, such as a cellular telephone, based on phase differences of
transmitted
signals of two or more frequencies received at two or more receiving sites.
Prior position detection, or navigation, systems may be generally divided into
two
categories: passive navigation systems and active navigation systems. In a
passive
navigation system, a mobile device determines its position based on signals
received from
transmitters positioned at known locations. In an active navigation system,
the mobile
device transmits signals which are received by one or more receivers
positioned at known
locations. The position of the mobile device is then determined based on the
received
signals and the known position of the receivers.
In the past, passive navigation systems have been generally favored over
active
navigation systems. One reason for this favoritism is that in active systems
each mobile
unit must transmit signals to determine its position. Since the number of
mobile units in
a single active system may be in the millions, these signals may overly
congest the active
system and may cause the system to malfunction. This problem is exacerbated in
situations.where the position of the mobile unit needs to be continuously
determined with
a high degree of accuracy. For example, an airplane needs virtually continuous
position
determinations due to its high velocity and, therefore, would need to transmit
signals
virtually continuously. Having' a large number of mobile units continuously
transmitting
such signals could possibly overload, or congest, an active navigation system.
One prior active navigation system was espoused by the GEOSTAR corporation.
In the GEOSTAR system, at least one orbiting satellite transmits signals to
mobile
transponders and receives replies from the mobile transponders. For
calculation
CA 02313485 2000-06-09
WO 99130181 PC"fNS98/24788
purposes, the GEOSTAit system assumes that the mobile transponder is on the
earth's
surface. By measuring the time a signal takes to travel from a satellite to a
mobile
transponder and back to the satellite (loop propagation delay), the mobile
transponder can
be determined to lie somewhere on a calculated sphere of appropriate radius.
Since the
intersection of the calculated sphere and the earth's surface is a circle, the
GEOSTAR
system thereby locates the mobile transmitter somewhere on the circle. If two
satellites
are employed, the mobile transponder can be deduced to also lie on a second
circle.
Since the intersection of these two circles is two points, the GEOSTAR system
can
therefore locate 'the mobile transponder at one of the two points.
The GEOSTAR system unfortunately exhibits some significant deficiencies.
Firstly, the GEOSTAR system suffers from the problem of congestion as
experienced by
other prior active navigation systems. Secondly, the GEOSTAR system relies
upon loop
delay measurements which are typically of questionable reliability or
accuracy. For
1 S example, the accuracy of loop delay measurements is deleteriously affected
by timing
errors in the transponder hardware which may be caused by any number of known
factors, such as temperature, imprecise manufacturing tolerances and the like.
In
addition, GEOSTAR systems typically use wideband signal transmissions to
assure
accurate loop delay measurements. Unfortunately, such wideband signal
transmissions
occupy a large portion of available bandwidth, thus significantly contributing
to the
aforedescribed problem of congestion.
Capacity is an issue at the heart of cellular communications systems and
satellite-
based mobile communications systems. Enough capacity for all users is
guaranteed by
dividing the service area into a large number of small cells with the ability
to re-use the
limited number of available radio frequencies again in different cells which
are
adequately sparated. U.$. Patent No. 5,619,503 issued to Dent on April ~8,
1997 describes
improvements to multi-cell or multi-beam communications systems that permit
higher
capacity by denser frequency re-use - ultimately permitting every frequency
channel to be
used for a different purpose in every cell or beam. The disclosure of U.S.
Patent No.
2
CA 02313485 2000-06-09
WO 99/30181 PCT/US98/24788
5,619,503 is hereby incorporated by reference herein and provides the capacity
improvements necessary to allow an active navigation system to succeed.
Methods to
obtain an initial coarse position estimate are also described and may be used
in the
current invention.
There is thus a need in the art for a system and method for providing active
position determination of a mobile transmitter which increases the accuracy of
the
position determination.
Summary of the Invention
This need is met by a method and system in accordance with the present
invention
wherein a first signal at a first frequency and a second signal at a second
frequency are
received from a transmitter at first and second receiving stations. Based on
phase shifts
of the signals received at the first and second receiving stations, a range
difference is
calculated which locates the transmitter on a first hyperbola having the
stations as foci.
In accordance with one aspect of the present invention, a method is provided
for
determining position of a mobile transmitter, such as a cellular telephone,
relative to first
and second receiving sites. Either of, or both of, the first and second
receiving sites may
be a land-based cellular base station or a satellite relay station. A first
signal at a first
frequency is transmitted by the mobile transmitter and received at the first
and second
receiving sites, as respective first and second received signals. A second
signal at a
second frequency is transmitted by the mobile transmitter and received at
first and second
receiving sites, as respective third and fourth received signals. A first
phase difference
measurement is made based on the first and second received signals. A second
phase
difference measurement is made based on the third and fourth received signals:
A
position of the mobile transmitter is determined based on the first and second
phase
difference measurements, the first and second frequencies and the first and
second known
locations. In particular, the f rst and second phase difference measurement
are scaled
3
CA 02313485 2000-06-09
WO 99/30181 PCT/IJS98/24788
based on the first and second frequencies and a range difference is
determined. For
example, a frequency difference may be determined between the first and second
frequencies and the difference between the first and second phase difference
measurements is scaled by the frequency difference to obtain~a range
difference.
The first phase difference measurement is preferably performed by determining
a
first phase shift for the first received signal; determining a second phase
shift for the
second received signal; and making the first phase difference measurement
based on the
first and second phase shifts. Similarly, the second phase difference
measurement is
preferably performed by determining a third phase shift for the third received
signal;
determining a fourth phase shift for the fourth received signal; and making
the second
phase difference measurement based on the third and fourth phase shifts.
The calculated range difference places the transmitter on a first hyperbola
with
1 S foci at the first and second receiving stations. To further define the
position of the
transmitter, the method may further comprise the steps of calculating third or
fourth phase
difference measurements for a third receiving site and one of the first or
second receiving
sites; localizing the transmitter on a second hyperbola related to the third
receiving site
and the one of the first and second receiving sites; and determining the
position of the
mobile transmitter based on the intersection of the first and second
hyperbolas.
The first frequency may be a random access channel and the second frequency
may be a traffic channel. Preferably, at least one of the first and second
frequencies is
contained in set of frequencies allocated to the mobile transmitter for
frequency hopping
communication. The signal on at least one of the first and second frequencies
may be
comprised of signal bursts formatted in accordance with a time division
multiple access
format.
One aspect of the present invention is the process for making the required
measurements. For example, the step of making a first phase difference
measurement and
4
CA 02313485 2000-06-09
WO 99/30181 PCT/US98/24788
the step of making a second phase difference measurement may comprise the
steps of
converting the first, second, third and fourth received signals into a digital
stream of
numerical values representative of instantaneous phases of the first, second,
third and
fourth received signals; and determining the position of the transmitter based
on the
stream of numerical values and the first and second frequencies:
Additionally, the step of converting may comprise the steps of downconverting
the first and third received signals to a first intermediate frequency;
downconverting the
second and fourth received signals to a second intermediate frequency; and
sampling and
analog-to-digital converting the downconverted first, second, third and fourth
received
signals. The first and second intermediate frequencies may be generated by
local
oscillators which may be synchronized to a common reference frequency. The
common
reference frequency could be derived from satellite navigation signals.
1 S If the intermediate frequency is set to approximately zero, the step of
sampling
and analog-to-digital converting may comprise the step of quadrature
converting the
downconverted first, second, third and fourth received signals to yield the
stream of
numerical values containing complex values. Alternatively, the step of
sampling and
analog-to-digital converting may comprise the step of logpolar converting the
downconverted first, second, third and fourth received signals to yield the
streaift of
numerical values containing numerical value pairs representative of
instantaneous signal
phase and amplitude.
The step of determining the position may include the step of determining
first,
second, third and fourth phase difference measurements for the first, second,
third and
fourth received signals. The stream of numerical values can then be correlated
to
determine an integral number of samples delay between arrival of the first and
second
received signals and the third,and fourth received signals and to determine
respective first
and second correlation phases. The correlation step may comprise determining
the
5
CA 02313485 2000-06-09
WO 99/30181 . PCT/US98/24788
correlation phases at the first frequency and the second frequency and using
the
correlation phases to determine the first and second phase difference
measurements.
The present invention further provides for translation in time of the phase
measurements. For example, the step of making a first phase difference
measurement
comprises the step of performing multiple measurements of the first phase
difference
measurement at different instants of time and interpolating or extrapolating
the multiple
measurements to determine a value for the first phase difference measurement
at an
instant of time at which the second phase difference measurement was made.
By varying the difference between the first and second frequencies in
accordance
with the present invention, accuracy can be adjusted. In particular, a
frequency difference
between the first and second frequencies is determined. The frequency
difference is then
compared to a predetermined frequency value. If the determined frequency
difference is
less than the predetermined frequency value, then a new value for at least one
of the first
and second frequencies is selected such that the frequency difference is
larger. The first
and second phase difference measurements are then repeated using the new value
for the
at least one of the first and second frequencies.
In accordance with another aspect of the present invention, a system foi
determining position of a mobile transmitter is provided. The mobile
transmitter is
capable of transmitting fast and second signals at respective first and second
frequencies.
The system comprises a first and second receiving station for receiving the
first signal
and a network control processor for computing a first phase difference based
on the first
signal received at the first and second receiving stations. The first and
second receiving
stations also receive the second signal and the network control processor
coputes a
second phase difference based on the second signal received at the two
stations. The
network control processor then determines a range difference measurement of
the mobile
transmitter based on the first and second phase difference measurements and
the first and
second frequencies. The first receiving station comprises a first phase
measuring circuit
6
CA 02313485 2000-06-09
WO 99/30181 PCTNS98lZ4788
for determining a first and third phase value for the first and second signals
respectively
as received at the first receiving station. Similarly, the second receiving
station
comprises a second phase measuring circuit for determining a second and fourth
phase
value for the first and second signals respectively as received at the second
receiving
station.
These and other features and advantages of the present invention will become
apparent from the following detailed description, the accompanying drawings
and the
appended claims.
Brief Description of the Drawines
Fig. 1 is a diagram of a system in accordance with the present invention
comprising a mobile transmitter and three receiving stations;
Fig. 2 is graphical representation of successive phase difference measurements
taken while the transmitter shown in Fig. 1 is frequency hopping;
Fig. 3 is a graphical representation of phase difference measurements made on
sequentially different frequencies;
Fig. 4 is a diagram showing non-simultaneous arrival of a signal at two of the
receiving stations shown in Fig. 1;
Fig. 5 is a diagram showing phase difference measurements between signals not
arriving at the same time at the two receiving stations;
Figs. 6a and 6b are graphical representations of an exemplary autocorrelation
functions of random digitally-modulated signals;
CA 02313485 2000-06-09
WO 99130181 PGTNS98/24788
Fig. 7 is a schematic of a system in accordance with the present invention
wherein
at least one of the receiving stations is a satellite; and
Fig. 8 is a schematic diagram of an exemplary dual channel receiving system in
accordance with one aspect of the present invention.
Detailed Description of the Invention
A system 100 for determining position of a mobile transmitter 102 in
accordance
with the present invention is shown in Figure 1. The mobile transmitter 102
may be a
cellular telephone, satellite communicator or any other mobile device which
transmits
signals. First, second and third receiving stations, or sites, 104, 106 and
108 located at
known positions receive signals transmitted by the mobile transmitter 102. The
receiving
sites 104, 106 and 108 may be land-based cellular base stations belonging to a
mobile
telephone network, may be orbiting satellite relay stations in communication
with at least
one land-based station, or combinations thereof. A network control processor
110 may
control the receiving stations 104, 106 and 108, receives information from the
receiving
stations 104, 106 and 108, and, based on this information, calculates the
position of the
mobile transmitter 102. The network control processor 110 may also provide a
phase or
frequency reference signal SR~to the receiving stations 104,106 and 108. It is
to be
understood that the receiving stations 104, 106 and 108 and the network
control processor
110 may be widely separated, thus signals are transmitted between the devices
over
communications links, such as radio links.
Although shown as a separate unit, the network control processor 110 consists
of
various circuits and processors for performing calculations. As those skilled
in the art
will readily comprehend, these circuits and processors may alternatively be
contained in
one of the receiving stations ~ 04, 106 and 108 or may be divided among a
plurality of
sites. The structure, control and arrangement of the conventional components
and circuits
have, for the most part, been illustrated in the drawings by readily
understandable block
8
CA 02313485 2000-06-09
WO 99/30181 PCTNS98lZ4788
representations and schematic diagrams. Only those specific details are shown
that are
pertinent to the present invention in order not to obscure the disclosure with
structural
details which will be readily apparent to those skilled in the art having the
benefit of the
description herein.
The present invention may be advantageously.implemented using two or more
receiving stations. In particular, the mobile transmitter 102 transmits a
first signal S, at a
first fi~equency w, which is received by the first and second receiving
stations 104 and
106, as respective first and second received signals. The first frequency w,
may be a
Random Access Channel (R.ACI-~ which is used when the transmitter 102 first
initiates
contact with the stations 104, 106 and 108 and the processor 110. As is known,
the first
signal S, will experience a first phase shift ~," during transmission to the
first receiving
station and a second phase shift ~8, during transmission to the second
receiving station
106. The magnitudes of the phase shifts c~~, and ~B, depend upon the distance
of travel of
the first signal S, to the respective stations 104 and 106. Consequently, the
first and
second phase shifts ~~, and ~B, will not be equal unless the mobile
transmitter 102 is
equidistant from the first and second receiving stations 104 and 106.
First and second receiving stations 104 and 106 receive a frequency and phase
reference signal from a common source, such as the GPS satellite system of the
network
control processor 110. The stations 104 and 106 measure the received phases
~~, and ~B,
of the first signal S, using the common phase reference signal, and pass the
phase
measurements to the network control processor.
The network control processor 110 determines a first phase difference
measurement ~~, in accordance with the following equation:
(1)
9
CA 02313485 2000-06-09
WO 99/30181 PCT/US98/24788
The mobile transmitter 102 next transmits a second signal Sz at a second
frequency wz which is similarly received by the first and second receiving
stations 104
and 106, as respective third and fourth received signals. The second frequency
wz may be
a traffic channel, for example, assigned to the transmitter 102 for passing
communications tragic after the initial contact on the R.ACH channel. The
mobile
transmitter 102 may change between, or transmit at, the first and second
frequencies w,
and wz in a predetermined frequency change schedule (frequency hopping
schedule), on
command from the network control processor 110 or other control mechanism. The
second signal Sz will experience a third phase shift ~,~ during transmission
to the first
receiving station 104 and a fourth phase shift fez during transmission to the
second
receiving station 106. The magnitudes of the phase shifts ~,~ and ~BZ will
depend upon
the distance of travel of the second signal Sz to the respective stations 104
and 106. In an
anlogous manner, the first and second stations 104 and 106 determine the
phases ~,,z and
Paz using the common reference signal and pass the measurements to the network
control
processor 110. The network control processor 110 then computes a second phase
difference Ocøz using:
D~z = ~~z - ~sz (2)
A time difference dT is next determined using the following equation:
dT=D -~~z
wt-wz
The time difference dT is a measure of the difference in propagation delay of
the
first and second signals S, and Sz to the respective first and second
receiving stations 104
and 106. Since this difference in the propagation delay is dependent upon the
difference
in the distances between each of the stations 104 and 106 and the mobile
transmitter 102,
the time difference dT is also a measure of the range difference between each
station 104
CA 02313485 2000-06-09
WO 99130181 PCT/US98/24~88
and 106 and the mobile transmitter 102. A range difference dR is calculated,
preferably
by the network control processor 110, using the following equation:
dR = c * dT (4)
S
wherein c is the speed of light (3 x 10$ m/s). The value (range difference
measurement)
calculated for the range difference dR places the location of the transmitter
102 on a first
hyperbolic curve contained in a family of hyperbolic curves having the first
and second
receiving stations 104 and 106 as foci.
Using the third receiving station 108, which is preferably not collinear with
the
first and second stations 104 and 106, the network control processor 110 can
receive
other measurements that locates the transmitter 102 on a second hyperbolic
curve. The
intersection of the first and second hyperbolic curves defining the position
of the
1 S transmitter 102. Alternatively, loop delay measurements, if available, can
be used to
determine the distances of the transmitter 102 from the first station 104, the
second
station 106 or both, thereby identifying a segment of the first hyperbolic
curve on which
the transmitter 102 is located.
The invention is envisaged to be most useful when loop delay measurements are
not available or not as accurate as the delta-range measurements made in the
above
described manner. In those circumstances, loop delay measurements may be used
only to
resolve ambiguities, although a position uncertainty will still be present in
the dimension
lying along the first hyperbolic curve. The position accuracy is much better,
however, in
the dimension perpendicular to the first hyperbolic curve, as this is
determined by the
more accurate delta-range measurement.
Various sources of error may cause inaccuracies in the calculation of the
delta-
range measurement dR. Since the transmitter 102 does not transmit at the first
and
second frequencies w, and w2 concomitantly, one source of error may arise if
the
11
CA 02313485 2000-06-09
WO 99/30181 PGTNS98/24788
transmitter 102 or any of the receiving stations 104, 106 or 108 are moving.
In such a
scenario, it is possible for the transmitter 102 to transmit at the first
frequency w, in one
location and then to have moved to another location before transmitting at the
second
frequency w2. The signal at the second frequency w2 would therefore have phase
shifting
characteristics based, in part, on the distance traveled from the second
location. One
solution to this problem, and most likely the optimum solution, is to use a
Kalman filter
algorithm to track changes in all parameters, as discussed more fully below.
Frequency errors is the common reference signal and any errors due to movement
of the'transmitter 102 or receiving stations 104, 106 and 108 may cause
successive
changes in the phase difference measurements ~~, and ~~2. A hypothetical
progression
of phase difference measurements taken at times ta, t,, tz, t3.... is shown in
Fig. 2. For ease
of description the phase difference measurements have been labeled ~,X and ~Zx
where the
subscripts "1" and "2" represent the first frequency w, and the second
frequency wz,
respectively, and the subscript "x" represents a variable integer indicating a
time the
measurement was taken. For example, a phase difference measurement ~,6 is a
measurement of a signal having the first frequency w, taken at the time tb.
For the example illustrated in Fig. 2, the transmitter 102 alternates, or
"hops",
between transmitting the first signal S, at the first frequency w, during even
times to, tz, t4
... and transmitting the second signal S~ at the second frequency w2 during
odd times t,, tj,
is ... The graph in Fig. 2 has further been drawn in a modulo-2~ format in
that the phase
difference measurements begin again at zero for every 2~ or 360°. As
can be seen, phase
difference measurements ~,~"~".X for the first frequency w, and phase
difference
measurements ~~_X for the second frequency w2 move progressively up sloped
lines.
The slope of the lines is not due to making the measurements at different
frequencies, as
all measurements on a line are made at either the first or second frequency w,
or w2.
Consequently, the slope mint be due to the sources of frequency error
mentioned above.
12
CA 02313485 2000-06-09
WO 99/30181 PGT/US98/Z4788
The sloped lines~representing the phase difference measurements ~~-x for the
second frequency w2 may have a slightly different slope than the sloped lines
representing
the phase difference measurements ~I~,~~.x for the first frequency wl. The
slight difference
in slope is due to the fact that the rate of change of phase caused by a fixed
frequency
error is constant, but the rate of change of phase due to a constant
percentage of
frequency error is proportional to the frequency. The difference in slope
between the two
lines will most likely be practically negligible, so that an average slope can
be determined
using the slope of both lines. Thus, the average slope can be determined by
averaging as
set forth below:
A= ~~z- to
t2 - '0
'Y 14 _ 'Y 12
t4 - t2
C Y'16 - 4
-
~_t4
D
=
t3 _ ti
E ~25-~23
=
t5 _ ~3
and;
F Y'27'Y25~2S
=
t, - t5 . . .
etc.
wherein the phase differences are computed modulo-2n (i.e. 2~ is added if the
difference
becomes negative). An average, or determined, slope dW can now be determined
by the
following equation:
dW= A+B+CH-D+E+F+...
(Number of slopes calculated)
13
CA 02313485 2000-06-09
WO 99/30181 PCT/US98JZ4788
Using the determined slope dW which has the dimensions of a frequency error,
values of
phase differences at the first frequency w, and the second frequency wz may be
computed
at substantially the same instant in time by interpolation.
Thus a phase measurement ~2, at the second frequency w2 referred to instant of
time t,
may be determined by:
- ~~o - dW*~tt - to)i
or,
~~~ - ~m - dW*(tx - t,)~
or the average of the two values,
~~o '~' ~u '~- dW*(2tt - tx - ~)
2
The latter simplifies to: o + ~
2
if the instants of time ta, t,, t~... are exactly equispaced in time, and the
effects of noise are
ignored.
Any prior art method of interpolation, such as least squares curve fitting,
may
alternatively be used to obtain a best estimated for the phase measurement
c~2,. Thus, the
value of the phase measurement at the second frequency w2 may now be
subtracted from
the phase measurement made at the first frequency w, at the same instant of
time t, to
obtain a difference value which is corrected for the slope error. This may
repeated at
other instants of time tz, t,. .. by interpolating either the line for phase
measurements at
the first frequency w, or the line for phase measurements at the second
frequency wz, as
appropriate, to obtain a succession of phase differences that depend only on
the frequency
change. If the transmitter 102 and the receiving stations 104, 106 and 108 are
knowri to
14
CA 02313485 2000-06-09
WO 99/30181 PGTIUS98I24~88
be stationary, these phase measurements must all correspond to the same range
difference
dR and the computed range differences at times t,, t2, t3.... may be averaged
to obtain
higher accuracy.
Alternatively, the successive range differences dR computed at times t,, tZ,
t,....
may show a systematic drift, revealing a possible movement of the transmitter
102 or the
receiving stations 104, 106 or 108 that was not previously modeled. The
principle of a
Kalman filter is to model all such movement in order to predict what the range
difference
should be at a given instant of time. For example, motion of the transmitter
102 may be
modeled by assuming a constant velocity which has to be estimated.
Alternatively,
motion of two receiving stations in earth orbit may be modeled by use of sets
of orbital
parameters, such as altitude, inclination, eccentricity and the like. When the
models
predict a range difference that differs from the value computed from
measurements at a
given instant, the Kalman procedure updates the model parameters such that, on
average,
1 S the error between the prediction and the measurement is minimized in a
least squares
sense. When the system of the present invention is used to determine the
positions of
large numbers of mobile transmitters, the parameters modeling the motion of
the
receiving stations are common, thus resulting in a very accurate determination
of the
motion of a receiving station (e.g. satellite orbits) based on receiving
signals from many
mobile transmitters. This in turn aids accurate position determination for
each individual
transmitter.
Figure 3 illustrates an alternative measurement sequence when frequency
hopping
between two or more frequencies is not employed, but other programmed
frequency
changes are made. In.Figure 3, the transmitter starts out transmitting at the
first
frequency w, and a signal phase difference at two receiving stations, such as
the first and
second receiving stations 104 and 106, is determined at times to, t, and t2 to
be phase
differencemeasurements ~"" ~", and ~,2 respectively. The first frequency w,
can, for
example, be the RACH channel frequency as explained above. The transmitter
then
changes to the second frequency wz which can be a first traffic channel
frequency having
CA 02313485 2000-06-09
WO 99/30181 PCT/US98/24788
a small frequency difference from the RACH channel frequency, for example
SOKHz.
Phase difference measurements are then made at the second frequency wz at
times t~, t4
and is to obtain phase difference measurements ~z3, ~z4 and ~zs. The
transmitter may
optionally then switch to another traffic channel frequency, or third
frequency w,, at times
S tb, t, and tg at which phase difference measurements ~j6, ~3, and ~3$ are
made.
It will be obvious from Figure 3 that phase difference measurements at the
first
frequency w, and the second frequency wz at the same instant in time, such as
tz or t,, may
be obtained by backward extrapolation of the slope line for the second
frequency wz
measurements or by forward extrapolation of the slope line for the first
frequency w,
measurements, or both. Thus, a first estimate for a first range difference dR,
is obtained.
Likewise, the slope line at the third frequency w, may be backwards
extrapolated and the
slope line at the second frequency wz may be forward extrapolated to obtain a
second
range difference which must be of the same order as the first range difference
dR,. If ~,~3 -
wz is much larger than wz - w,, there will be 2~ ambiguities. This 2~
ambiguity may be
resolved by adding or subtracting multiples of the wavelength at the larger of
the higher
difference frequency, that is adding or subtracting by multiples of 2~*c/(w3 -
wz) until the
range difference lies close to the first estimate for the first range
difference dR,. The
second estimate for the second range difference dRz will however, be more
accurate than
the first estimate for the first range difference dR,. Implementing other
methods of
processing the data collected according to Figure 3, such as best fitting
straight lines in a
least squares sense, or Kalman filtering, is well within the capabilities of a
person skilled
in the art.
The above description assumes that the transmitter 102 transmits signals that
are
received at two separated receiving stations, the receivers then being able in
some way to
convey the received signals to a common point for phase comparison. Making a
phase
comparison implies that a signal from one station is available at the same
time as a signal
from another station.
16
CA 02313485 2000-06-09
WO 99/30181 PCT/US98/24788
However, a preferred waveform to be transmitted by the transmitter for
communications purposes is a short signal burst occupying a timeslot of a
repetitive
TDMA frame period. In satellite systems, it is very possible that the
differential delay
between to paths exceeds the duration of a timeslot, so that the signal is not
received at
S the first receiving station 104 and the second receiving station 106 in
overlapping time
periods. In this case, a different method of determining phase difference
which does not
rely on the signals from the stations 104 and 106 being present at both inputs
of a phase
comparator simultaneously is required.
This situation is depicted in Figure 4. It may be seen that there is no
overlap
between the period during which a signal burst transmitted from the
transmitter 102 is
received at the first station 104 and the period during which the same signal
burst is
received at the second station 106. This occurs when the range difference dR,
which is
equal to RANGE (1) - RANGE (2) in Figure 4, exceeds the velocity of light c
times the
1 S burst duration. There is of course no problem in measuring the phase
difference of arrival
when the signal is a continuous wave signal (CW) which would be present at
both
receiving stations 104 and 106 simultaneously. The invention thus comprises
the
optional implementation of commanding the transmitter 102 to transmit a CW
signal for
a sufficient duration to guarantee that the CW signal arrives at least two
receiving stations
in overlapping time periods. The two received signals may then be applied to
the two
inputs of a phase comparator which determines the phase difference of arrival
during the
overlapping time period when both signals are present.
If there is no overlapping time period when both signals are present, it is
obviously fruitless to apply the signals to the inputs of a phase comparator.
Consequently, a novel solution is clearly required for measuring the phase
difference of
arrival between two signals that do not arrive at the two stations in an
overlapping time
period.
17
CA 02313485 2000-06-09
WO 99/30181 PCTNS98/Z4788
Figure 5 illustrates the principle of such a solution. The transmitter 102
transmits
a signal burst which is received at first and second antennas 112 and 114 of
the respective
first and second receiving stations 104 and 106. The respective received
signals are
downconverted in first and second receivers ItX-A 116 and RX-B 118 to a
suitable
frequency for analog-to digital (Atop) conversion and sampling. For example, a
suitable
frequency spectrum for Atop conversion and sampling is the complex baseband.
In the
complex baseband, a signal is mixed with Cosine and Sine oscillator signals to
produce
In-Phase and Quadrature (I, Q) signals that each extend in the spectral domain
from zero
(DC) to half the receiver bandwidth. The I and Q signals are then sampled and
digitized
by sampling first and second Atop convertors 120 and 122 in which it is
understood that
each comprises two conversion channels, one for the I signal component and one
for the
Q signal component. The resulting first and second signal sample streams A,,
Az, A3...
and B,, BZ, Bj... are complex samples having both a real (I) part and an
imaginary (Q)
part. An alternative technique for digitizing a signal while preserving its
complex vector
representation is the LOGPOLAR technique described in U.S. Patent no.
5,048,059
issued to Dent, which is hereby incorporated herein by reference.
In order to preserve the relative phase-of arrival information, it is
necessary to
employ oscillator synchronizing means 124 to synchronize all local oscillators
used in the
first and second receivers RX-A 116 and RX-B 118 so that the effects on signal
phase of
the downconversion processes are at least known. Likewise, it is necessary to
employ
convertor synchronizing means 126 to synchronize the sampling of the first and
second
Atop convertors 120 and 122 such that their sampling instants are the same or
at least
have a known relationship. The oscillator and converter synchronizing means
124 and
126 can, for example, comprise atomic clocks located at the first and second
receiving
stations 104 and 106 having stabilities of one part in 10 to the power 13, or
can comprise
GPS satellite navigation receivers which provide accurate time and frequency
references.
The GPS receivers may also be used to correct slow drifts of the atomic
clocks. The GPS
receivers are in any case desirable in order to accurately determine the
locations of the
first and second receiving antennas 112 and 114, as accurate knowledge of the
position of
18
CA 02313485 2000-06-09
WO 99130181 PCTN898/24788
signal reception is needed to compute the location of the transmitter 102.
Assuming,
therefore, that the oscillator and convertor synchronizing means 124 and 126
maintain all
local oscillators arid sampling clocks in phase and in time, all the
information on relative
time and phase of arrival of the transmitter signal at the antennas 112 and
114 is
contained within the complex first and second signal sample streams A,, AZ,
and A3.. .
and B,, Bx, B3...
The signal burst from the transmitter 102 affects sample numbers A"...A.~~~~
generated from the first receiver RX-A 116 but, due to delay difference,
affects different
sample numbers B,~...Bt~tL~ generated from the second receiver RX-B 118. A
conventional crosscorrelation between the first signal sample stream Ax and
the second
signal sample stream Bx is defined by:
i-L
C(k) = 1 /L ~ E ( A(j + i) ~ B * (j + k + i)~
3 ~~ L
where * means complex conjugate and C(k) is the correlation value. The above
computation is repeated for many consecutive values of 'k' to obtain a set of
correlation
values C(1), C(2).....etc. The values of 'k' should range around the nominal
time
difference of arrival of m-n samples from a smallest possible delay value to a
largest
possible delay value.
The correlation value C(k) represents the similarity of the first sample
stream A
starting at sample A(j+1) to the second sample stream B delayed by 'k'
samples, i.e.
starting at sample B(j+k+1). The similarity is determined over an interval of
L
consecutive signal samples. The correlation value C(k) is a complex number
that has a
correlation magnitude and a correlation phase. The correlation magnitude,
plotted as a
function of 'k', exhibits a characteristic curve, termed the autocorrelation
function, which
is determined by the modulation pattern imposed on the signal.
19
CA 02313485 2000-06-09
WO 99/30181 PCTNS98/24788
If the signal is a random digital modulation pattern of sufficient length, or
is
deliberately chosen to be one of certain sequences that produce desirable
autocorrelation
functions, then the autocorrelation function is as depicted in Figure 6.
Figure 6a shows
the classical form of the autocorrelation function for a digital signal that
is not restricted
in bandwidth. The magnitude peaks at unity when the signals at the inputs of
the
correlator are perfectly aligned, and falls to zero at plus and minus one
digital symbol
period away from the peak. A digital symbol period is equal to a bit period if
the
modulation is binary phase shift keying (BPSK) or equal to two bit periods if
the
modulation is Quadrature Phase Shift Keying (QPSK). In practice, modulation
signal
bandwidths and receiver bandwidths are restricted to prevent adjacent channel
interference. More generally, the Wiener-Khintchine theorem states that the
autocorrelation function is the inverse Fourier transform of the signal
spectrum, which is
affected by transmitter and receiver filtering. As depicted in Figure 6b, a
typical,
practical autocorrelation function is more rounded than the function shown in
Figure 6a
and oscillates with decreasing amplitude away from the peak.
When the magnitude of the autocorrelation function is plotted, the negative
swings of the function will be reflected into the positive half plane (shown
by the dashed
lines in Figure 6b.). The width of the main lobe between the first.nulls is
approximately
equal to the reciprocal of the signal's main spectral lobe. For example, if
the signal
spectral lobe is of the order of 40KHz wide, then the autocorrelation peak
will be of the
order of 25 microseconds wide. Thus a first coarse delay estimate of the time
difference
of arrival may be obtained by correlating the first complex signal stream A
with the
second complex signal stream B and determining for which value of the sample
delay 'k'
the correlation magnitude is maximum. The accuracy of this determination will
be some
fraction of 25 microseconds, for example, one tenth of 25 microseconds or
~2.SuS.
The coarse delay.estimate was obtained from the magnitude of the correlation
value C(k) ignoring the phase. The phase of the correlation value C(k) allows
a much
CA 02313485 2000-06-09
WO 99/30181 PCTNS98/2478$
more accurate determination of the delay however. A more mathematical
formulation
will illustrate the value of phase.
Let the transmitted signal, designated by S(t), be given by the equation:
s(t) = z(t)EXP~wr)
where Z(t) is a complex function describing both phase and amplitude
modulation (if not
a pure phase modulation or a pure amplitude modulation) of the signal's
carrier
frequency, and EXP(jwt) represents the unmodulated carrier frequency. The
transmitted
signal S(t) is received with a first delay TI at the first receiving station
104 and with a
second delay T2 at the second receiving station 106. Signals A(t) received at
the first
receiving station 104 and signals B(t) received at the second receiving
station 106 are
thus given by:
A(t) = Z(t-T 1 )EXP(j w(t-T1 ))
and, B(t) = Z(t-T2)EXP(jw(t-T2))
The signals A(t) and B(t) are sampled at instants of time t,, t~, t3....t; =
i:dT, where dT is
the interval between samples assuming synchronization at both stations 104 and
106.
Thus giving rise to the first and second signal sample streams A(i) and B(i)
given by:
A{i) = A(i~dT) = Z(i~dT-T1)EXP(jw(i~dT-T1))
B(i) = B(i~dT) = Z(i~dT-T1)EXP(jw(i~dT-T2))
Then the correlation value C(k) can be expressed as shown:
i=L
C(k) = 1/L E ( Z[(j + i) ~ dT - T1] ~ Z[(j + k + i)dT - T2] ~ EXP[jw(k~dT -
(T1 - T2)]]
i=1
21
CA 02313485 2000-06-09
WO 99130181 PCT/US98124788
Since the complex exponential term is independent of the index of summation
'i', it may
be moved outside the summation, obtaining:
i=L
Then C(k) = 1/L EXP[jw(k~dT-(T1-T2)]~E[Z[(j+i)~dT-Tl]~Z((j+k+i)dT-T2]]
i=1
= Cm(k)~EXP[jw(kdT-(Tl-T2)]
where Cm(k) is an autocorrelation function of the signal modulation Z(t)
detached from
the cagier frequency 'w' , and whose autocorrelation function was exemplified
in Figure
6b.
Since the complex exponential term is always of unit magnitude, it does not
affect the
magnitude of the correlation value C(k); which is equal to the autocorrelation
function
Cm(k), but does affect its phase angle by the amount given by:
~(k) = w(k~dT-(T1-T2))
In finding the correlation value ~C(k) with maximum magnitude, we have found
the value
of 'k' for which k~dT is closest to the true delay difference Tl-T2, thus
minimizing c~(k).
Of course, ~(k) is, in any case, reduced modulo-2~, so it is not possible to
tell exactly
how many whole cycles of the carrier frequency w are missing. This corresponds
to an
ambiguity in the range difference measurement dR of an unknown number of
wavelengths of the carrier frequency w. With an error in the coarse estimate
of ~2.SuS
and a cagier frequency of 2GHz, the ambiguity is ~S00 wavelengths. It is not
possible,
therefore, to tell which one of 1,OOO.possible values is the true delay value.
This
ambiguity is resolved according to the invention by repeating the correlation
when the
transmitter is using a different frequency. For description purposes, assume
that the
about example was performed with a carrier frequency w equal to the first
frequency w,
and the different frequency is the second frequency w2. Thus, we obtain
~(l,k) = w,~(k~dT-(Tl-T2))
22
CA 02313485 2000-06-09
WO 99/30181 PCTNS98/24788
~(2,k) = wz~(k:dT-(Tl-T2))
with care being taken at the receiving stations to compare correlations having
the same
value of 'k'.
On the assumption that the range difference dR = Tl-T2 does not change between
the instants of time that the transmitter transmits on the first frequency w,
and the instants
of time it transmits on the second frequency w2, the above phase values are
subtracted to
give a phase difference dc~:
d~ ° ~(l,k)-~(2,k) _ (u'~ - ~'~)~(k~dT-(Tl-T2))
It will be recalled that the value of k~dT-(Tl-T2) has been reduced by
choosing a value
for k on the order of ~2.SuS. By now choosing values for the first and second
1 S frequencies w, and w2 such that their difference w, - w2 is less than
~/SuS (i.e. the
frequency difference is 100KHz), for example, the value of the phase
difference d~
determined above will lie between -tn/2 and thus not be ambiguous. We may then
determine a more redefined value for the range difference Tl-T2 as follows:
T1-T2 = k~dT -
N'm ~'~
Suppose for example that w,-w2 is 2~ x SOKHz, and that we can measure the
phase
difference d~ with an accuracy of +5 degrees. Then residual error s~ on the
range
difference T1-T2 will be:
s~ _ + 5 / 360.50000 = ~ 0.27 pS
The accuracy has thus been improved by almost an order of magnitude from 2.5
~S to
0.27 p,S accuracy.
23
CA 02313485 2000-06-09
WO 99/30181 PCTIUS98/24788
The measurements can now be repeated with a greater value of the frequency
difference w,-wi. As we start now with a smaller uncertainty of 0.27 p,S which
allows
frequency differences of the order of IMHz to be used without creating modulo-
2~
ambiguities. Using IMHz frequency difference, the accuracy may be improved by
a
further factor of 20 to ~0.0135 ~S which represents a range difference error
of ~4 meters.
If the value of the range difference TI-T2 is not constant between the period
of
transmission at the first frequency w, and the period of transmission at the
second
frequency w2, the rate of change can be determined by the method illustrated
graphically
in Figures 2 and 3, namely by determining the difference between successive
measurements on the same frequency and then interpolating or extrapolating to
determine
the difference between measurements on different frequencies translated to the
same
instant of time.
It has thus been shown above how a combination of time-of arrival difference
measurements and complex correlation phase difference computations can be used
to
determine accurate range differences, and hence determine~the position of a
transmitter by
solving known hyperbolic navigation equations.
Figure 7 shows the use of first and second orbiting satellites 128, 130 as the
at
least two separated receiving stations needed by the invention. A signal from
the mobile
transmitter.102 is received at the satellites 128 and 130 and, after
translation of the signal
from the transmitter's original frequency to a feederlink frequency, the
satellites 128 and
130 relay the signal to respective first and second ground antennas 132 and
134.
Preferably, the antennas I32 and 134 are located at approximately the same
site on the
ground, thus facilitating connection to a common, dual-channel receiving
system 136.
Figure 8 shows more detail of an exemplary dual-channel receiving system
having
first and second receiving channels. The signal from the first antenna 132 is
fed to a first
24
CA 02313485 2000-06-09
WO 99/30181 PCT/US98I24788
downconvertor 201 and the signal from the second antenna 134 is fed to a
second
downconvertor 202. The first downconvertor 201 comprises a low noise amplifier
201 a,
a filter 201b, a mixer 201c and a local oscillator 201d. The second
downconvertor 202 is
identically constructed with a low noise amplifier 202a, a filter 202b, a
mixer 202c and a
local oscillator 202d. The downconvertors 201 and 202 preferably reside at the
focus of
their respective antennas since it is easier to transmit an intermediate
frequency output to
further processing circuits using flexible coaxial cable than to transmit the
original
microwave feederlink signal using a waveguide.
The intermediate frequency outputs from the downconvertors 201 and 202 are
further amplified and filtered in IF filter amplifiers 203 and 204. The
outputs of the IF
filter amplifiers 203 and 204 are fed to respective quadrature downconvertors.
Quadrature downconvertors 205 and 207 are supplied the signals received from
the first
satellite 128 and quadrature downconvertors 206 and 208 are supplied the
signals
received from the second satellite 130. The quadrature downconvertors 205,
206, 207
and 208 mix their input signals with a cosine and a sine wave signal from a
third local
oscillator 210 which is common to both receiver channels: The quadrature
downconvertors 205 and 206 output I-waveforms to respective
demultiplexer/digitizers
211 and 212. The quadrature downconvertors 207 and 208 output Q-waveforms to
the
respective demultiplexer/digitizers 211 and 212. The operation of a time-
multiplexed
feederlink is described in more detail in U.S. Patent Nos. 5,596,961,
5,555;271,
5,619,503, 5,619,210, 5,574,967, 5,568,088 and 5,642,358, the disclosures of
which are
hereby incorporated by reference herein.
The feederlink signal comprises a time multiplex of signals from the
satellite's
receiving antenna elements, together with some known test or calibration
signals. The
calibration signal may for example be set on board the satellite to 1+j0 so
that, when its
sample occurs in the time-multiplexed feederlink stream, it is known that the
I-waveform
value should be unity and the Q-waveform value should be zero. Demultiplexer
units
211 and 212 thus extract calibration signal samples from the I and Q waveforms
received
CA 02313485 2000-06-09
WO 99/30181 PCT/US98/24~88
from the quadrature downconvertors 205, 206, 207 and 208. These calibration
signal
samples are provided to automatic frequency control (AFC), automatic gain
control
(AGC) and synchronization (SYNC) units 213 and 214 where the calibration
samples are
compared with expected values and error signals produced. The error signals
are used for
three purposes: (1) to apply AGC to the IF filter amplifiers 203 and 204 such
that the
magnitude of the calibration samples is controlled to the expected value; (2)
to apply
AFC or automatic phase control (APC) to the first and second local oscillators
201 d and
202d such that the phase of the calibration samples is controlled to the
expected value;
and (3) to correct any sampling timing error in the demultiplexers 211 and 212
such that
the calibration samples are sampled in the middle of their optimum sampling
range. In
this way, all phase and amplitude matching errors in the two feederlink
channels from the
satellite through respective ground processing channels is removed up to the
outputs of
the demultiplexers 211 and 212.
The outputs of the demultiplexers 211 and 212 each represent the composite of
all
signals received by one of the satellite receiving antenna 132 or 134, which
may comprise
many mobile transmitter signals operating on different channel frequencies.
Digital
channel splitting units 215 and 216 first split the signal from each antenna
element into a
number of signals from each element corresponding to different channel
frequencies w,,
w2....Wn.. The set of antenna element signals.for a particular frequency, for
example w,, is
then fed to a digital beamformer, shown as reference number 217 for the first
satellite 128
and reference numeral 218 for the second satellite 130. The set of antenna
element
signals for frequency w" are fed to digital beamformers 219 and 220. Although
only four
beamforrners 217, 218, 219 and 220 are shown, there is a separate pair of
digital
beamformers for each frequency w,, w2 . . . w~.
The digital beamformers 217, 218, 219 and 220 combine the signals from each
satellite antenna element using complex coefficients in order to create
multiple, directive
receiving beams that discriminate signals from different directions. Although
the
exemplary system first performs channel splitting followed by direction
splitting, if is to
26
CA 02313485 2000-06-09
wo r~r~oia~ rcrius9sri4~ss
be understood that this is merely a preferred arrangement. The operations of
channel
splitting and direction splitting can be reversed by use of a wideband beam
former
followed by channel splitting on a per-beam basis. The advantage of the
preferred
arrangement is that it allows staggering of beam directions on different
channel
S frequencies (or even timeslots on a single frequency) in order to obtain
higher spectral
efficiency by greater frequency reuse on the ground, as is more fully
described in the
incorporated references.
When the transmitter 102 is located in one of the receiving beams so formed,
its
signal emerges enhanced by the appropriate one of the beam outputs of the
beatnformer
assigned to that transmitter frequency. For example, when the transmitter 102
is
transmitting on th,e first frequency, or channel w,, its signal emerges from
one beam
output of the beamformer 217 and a different beam output from the beamformer
218, as
the beam number is not necessarily the same for the first and second
satellites 128 and
130 as they illuminate the earth slightly differently because of their
separation in space.
Thus, the complex number streams A,, AZ, A,.... and B,, BZ, B3... depicted in
Figure S
emerge from the beamformers 217 and 218 respectively when the transmitter 102
is
operating on the first frequency w,, or from the beamformers 219 and 220 if
the
transmitter 102 is operating on the nth frequency w". It may be assumed that
the position
of the transmitter 102 is already known by some other means to an accuracy of
at least a
fraction of a beam's diameter on the ground, so that the beam output from
which the
transmitter signal emerges is known. This may for example be determined by
detecting a
Random Access Burst from the transmitter 102 in a RACH detector (not shown)
for that
beam, and not in the RACH detectors of other beams.
As well as processing sample streams A,, A~, A,... and B,, B2, B3... to
determine
the location of the transmitter 102, the streams A,, A2, A3... and B,, B2,
B3... may be fed
to a diversity demodulation and decoder to extract transmitted information, as
described
in U.S. patent application serial no. 08/305,904, entitled "Diversity-Oriented
Channel
27
CA 02313485 2000-06-09
WO 99/30181 PCTNS98/Z4788
Allocation In A Mobile Communications System", to Dent and Ewerbring and filed
on
December 12, 1994, the disclosure of which is hereby incorporated by
reference.
Once the sample streams A~, Ai, A,... and B,, B2, B3... are generated, they
may
be stored, or certain characteristics of them, such as correlations, rnay be
computed and
stored for off line processing. Thus, the determination of a transmitter's
location does
not need to be done in real time. The position can be deduced later, as
required, by
processing the stored sample stxeams retrospectively. Moreover, the invention
can
comprise using a number of fixed transmitters on the ground having accurately
known
positions to assist in determining necessary parameters such as satellite
position at a
given time. Signals from such transmitters may be transmitted periodically
and, after
emerging from appropriate channel/beam outputs of the ground processing, can
be stored
for processing off line.
The elements of the invention comprise transmitting a signal from the
transmitter
to be localized and receiving the signal at two separated receiving sites, the
signal being
downconverted, sampled and digitized at each site using synchronized or
otherwise
known local oscillator frequency and sampling time references. The respective
signal
samples are then communicated to a common location, such as a network
processor,
where they are processed to yield an accurate range difference from the
transmitter to the
respective receiving sites.
In one implementation, the two separated receiving sites comprise two orbiting
satellite relay stations that relay the received signals to a ground station.
In relaying the
signals, the satellites change the frequency received from the transmitter to
a different
frequency called the feederlink, on which the relayed signals are transported
to the
ground station. In that case synchronization means 124 shall comprise means to
synchronize the local osciliators on board respective spacecraft so that any
phase
difference therebetween is preferably zero, or at least predictable. Such
means can
comprise transmitting a pilot signal from the ground station to both
satellites, which may
28
CA 02313485 2000-06-09
WO 99/30181 PCT/US98/?,4788
be precompensated for Doppler shift separately for each satellite in order to
account for
satellite movement. The above listed patents, which were incorporated herein
by
reference, describe satellite transponder means which, in connection with
mufti-beam or
phased array satellite receiving antennas, allow the received phases of all
signals received
at each satellite to be preserved over the feederlink transmission, which
property is useful
in the context of this invention.
The invention described above with the aid of the incorporated references and
Figures 1 through 6 inclusive permits of a variety of implementations using
either land-
based stations or airport or orbiting satellite stations, and with a variety
of transmitters
using either CW signals or signals with arbitrary modulation, including short
TDMA
signal bursts. All such variations that may be made by a person skilled in the
art with the
aid of the above teachings lie within the spirit and scope of the invention as
described in
the following claims.
29
